Prompt2tree — AI skill for Claude Code

🌲 prompt2tree (`pt`)

[](https://pypi.org/project/prompt2tree/) [](https://pypi.org/project/prompt2tree/) [](https://opensource.org/licenses/MIT)

**The AI-native project scaffolder.** Convert raw LLM text trees into physical directory structures and initial boilerplate files in milliseconds.

---

💡 Why prompt2tree?

When using AI coding assistants (ChatGPT, Claude, Cursor, DeepSeek), asking agents to create individual folders and files burns thousands of context tokens and introduces latency or rate-limit delays.

**`prompt2tree` solves this.** Request a visual folder tree from your LLM, copy the string, and run `pt -c` (or `pt structure.txt`). Your entire project architecture is scaffolded instantly on your local machine for zero token cost.

---

⚡ Quick Start

1. Installation

pip install prompt2tree

*To enable direct OS clipboard reading, install with the `clip` extra:*

pip install "prompt2tree[clip]"

---

2. Usage Workflows

Option A: Directly from Clipboard (Fastest)

Copy any tree output from ChatGPT or Claude, then run:

pt -c
# or
prompt2tree --clip

Option B: From a Text File

pt structure.txt

*If no arguments are passed, `pt` automatically looks for `structure.txt` in the working directory.*

Option C: Piping via Stdin

cat structure.txt | pt

---

✨ Key Features

• 🧹 LLM Output Sanitization: Automatically strips Markdown backticks (```), conversational chatter, and inline explanatory comments (main.py # application entrypoint).
• 🌲 Universal Tree Support: Parses Unicode box-drawing trees (├──, └──), ASCII pipe/hyphen variants (|--, \--), and plain tab/space-indented hierarchies.
• 📄 Smart Boilerplate Auto-Fill: Automatically populate